There’s a lot riding on farmers’ ability to fight weeds, which can strangle crops and destroy yields. To protect crops, farmers have two options: They can spray herbicides that pollute the environment and harm human health, or they can hire more workers.
Unfortunately, both choices are becoming less tenable. Herbicide resistance is a growing problem in crops around the world, while widespread labor shortages have hit the agricultural sector particularly hard.
Now the startup FarmWise, co-founded by Sebastien Boyer SM ’16, is giving farmers a third option. The company has developed autonomous weeding robots that use artificial intelligence to cut out weeds while leaving crops untouched.
The company’s first robot, fittingly called the Titan — picture a large tractor that makes use of a trailer in lieu of a driver’s seat — uses machine vision to distinguish weeds from crops including leafy greens, cauliflower, artichokes, and tomatoes while snipping weeds with sub-inch precision.
About 15 Titans have been roaming the fields of 30 large farms in California and Arizona for the last few years, providing weeding as a service while being directed by an iPad. Last month, the company unveiled its newest robot, Vulcan, which is more lightweight and pulled by a tractor.
“We have growing population, and we can’t expand the land or water we have, so we need to drastically increase the efficiency of the farming industry,” Boyer says. “I think AI and data are going to be major players in that journey.”
Finding a road to impact
Boyer came to MIT in 2014 and earned masters’ degrees in technology and policy as well as electrical engineering and computer science over the next two years.
“What stood out is the passion that my classmates had for what they did — the drive and passion people had to change the world,” Boyer says.
As part of his graduate work, Boyer researched machine learning and machine vision techniques, and he soon began exploring ways to apply those technologies to environmental problems. He received a small amount of funding from MIT Sandbox to further develop the idea.
“That helped me make the decision to not take a real job,” Boyer recalls.
Following graduation, he and FarmWise co-founder Thomas Palomares, a graduate of Stanford University whom Boyer met in his home country of France, began going to farmers’ markets, introducing themselves to small farmers and asking for tours of their farms. About one in three farmers were happy to show them around. From there they’d ask for referrals to larger farmers and service providers in the industry.
“We realized agriculture is a large contributor of both emissions and, more broadly, to the negative impact of human activities on the environment,” Boyer says. “It also hasn’t been as disrupted by software, cloud computing, AI, and robotics as other industries. That combination really excites us.”
Through their conversations, the founders learned herbicides are becoming less effective as weeds develop genetic resistance. The only alternative is to hire more workers, which itself was becoming more difficult for farmers.
“Labor is extremely tight,” says Boyer, adding that bending over and weeding for 10 hours a day is one of the hardest jobs out there. “The labor supply is shrinking if not collapsing in the U.S., and it’s a worldwide trend. That has real environmental implications because of the tradeoff [between labor and herbicides].”
The problem is especially acute for farmers of specialty crops, including many fruits, vegetables, and nuts, which grow on smaller farms than corn and soybean and each require slightly different growing practices, limiting the effectiveness of many technical and chemical solutions.
“We don’t harvest corn by hand today, but we still harvest lettuces and nuts and apples by hand,” Boyer says.
The Titan was built to complement field workers’ efforts to grow and maintain crops. An operator directs it using an iPad, walking alongside the machine and inspecting progress. Both the Titan and Vulcan are powered by an AI that directs hundreds of tiny blades to snip out weeds around each crop. The Vulcan is controlled directly from the tractor cab, where the operator has a touchscreen interface Boyer compares to those found in a Tesla.
With more than 15,000 commercial hours under its belt, FarmWise hopes the data it collects can be used for more than just weeding in the near future.
“It’s all about precision,” Boyer says. “We’re going to better understand what the plant needs and make smarter decisions for each one. That will bring us to a point where we can use the same amount of land, much less water, almost no chemicals, much less fertilizer, and still produce more food than we’re producing today. That’s the mission. That’s what excites me.”
Weeding out farming challenges
A customer recently told Boyer that without the Titan, he would have to switch all of his organic crops back to conventional because he couldn’t find enough workers.
“That’s happening with a lot of customers,” Boyer says. “They have no choice but to rely on herbicides. Acres are staying organic because of our product, and conventional farms are reducing their use of herbicides.”
Now FarmWise is expanding its database to support weeding for six to 12 new crops each year, and Boyer says adding new crops is getting easier and easier for its system.
As early partners have sought to expand their deployments, Boyer says the only thing limiting the company’s growth is how fast it can build new robots. FarmWise’s new machines will begin being deployed later this year.
Although the hulking Titan robots are the face of the company today, the founders hope to leverage the data they’ve collected to further improve farming operations.
“The mission of the company is to turn AI into a tool that is as reliable and dependable as GPS is now in the farming industry,” Boyer says. “Twenty-five years ago, GPS was a very complicated technology. You had to connect to satellites and do some crazy computation to define your position. But a few companies brought GPS to a new level of reliability and simplicity. Today, every farmer in the world uses GPS. We think AI can have an even deeper impact than GPS has had on the farming industry, and we want to be the company that makes it available and easy to use for every farmer in the world.”
